Workshops, coaching, and strategy — from first automations to agentic software engineering to AI product strategy.
Agentic software engineering that deserves its name.
Plain language on the surface, production scars underneath. Start where you are; the next level is the same conversation.
“I keep hearing AI could save me hours, but I don’t know where to start — and half of what I try doesn’t stick.”
A working first automation and a clear-eyed map of what AI can and can’t do for your work.
Best first → half-day workshop · coaching hour
Start a conversation“We’ve automated a few things, but connecting AI properly into our actual tools and data is where it falls apart.”
A practical, guided path to connect AI into your real systems and workflows — without hiring a software team.
Best first → workshop → project consulting
Start a conversation“AI writes a thousand lines an hour. Left to vibe-code, most of it turns into spaghetti — duplicated, incoherent, impossible to change a year later.”
Keep AI fast and coherent: clean architecture, no redundancy, code that's still changeable a year from now — agents working like a disciplined engineering team, not a slop machine.
Best first → developer workshop track · team coaching
Start a conversation“Everyone’s asking what we should do with AI — in the company and in the product — and we need a real answer, not a trend deck.”
A prioritized map of where AI creates leverage (and where it doesn’t), and a scoped path from idea to pilot.
Best first → AI Opportunity Diagnostic → strategy workshop
Start a conversationHackathon coaching for any lane: take a team from idea to a working prototype, then decide what’s worth piloting, killing, or building for real.
The demo is easy. The version that survives real work is the job.
Most AI advice is a trend deck. I'm the opposite: a lawyer who learned to code at nine, an entrepreneur of 26 years, and an operator who ships production AI every day — not someone renting out opinions.
AI becomes useful when it's wrapped in discipline: staged systems, source anchoring, review, reversibility. That's what I teach, coach, and build.
Every engagement starts from a real problem and ends with something you can use. Start where it fits — each one opens the same conversation.
Different work for different teams — priced as different things, not the same day at two prices.
All prices exclude VAT & travel.
Not prompt tricks, not trend decks — capability grounded in systems that run in production.
Founded Uberchord in 2014; acquired by Duolingo in 2024.
Real production work, with AI agents running in parallel — the contribution graph above is a year of it.
Engaged with an intelligent-document-processing company.
Tests, gates, and judgment on every change. Speed without a lower bar for correctness.
Outputs are backed by the source or rendered explicitly empty — never confidently invented.
Decisions that carry real consequence keep a person in the loop. Some things shouldn’t be automated.
Changes that reach production can be rolled back. The principle is visible; the method is the work.
or email eckart@burgwedel.info